Relay advice please

[newshound]

No, it will switch AC or DC from an AC signal.

[newshound]

I shifted an application from octal base conventional relays to these because of relay failures, and these have been far more reliable (although not completely infallible).

[williamwright]

Come to think of it so have I. They switch quite frequently.

Bill

[The Natural Philosopher]

Well it still is not clear as to what the relay connects to what. But if its is just a matter of sensing mains and shoving a digital voltage to a pin then that opto isolator I linked too is perfect

[Martin Brown]

You might be better off and safer using a 12v PIR device then.

I presume that the trigger action is to pull the trigger input to the camera low in order to activate it. *How* are the relays failing?

Reviews online suggest some dodgy imitations/fakes about - where the coil burns out if they are on for significant periods of time..

[Brian Gaff (Sofa]

Yes snubber networks can stop a lot of that sort of thing and they won't make a big pulse of interference then either. Brian

[Brian Gaff (Sofa]

Still not clear, would not simply open circuiting the 12v supply do the job why are you earthing it or do you really mean, disconnect and earth the supply pin on the camera? Brian

[Brian Gaff (Sofa]

If you want sa 240 v relay, then simply strip down a wall wart 5v usb device for its pcb, and use a 5v relay solid stat or whatever on its output. Brian

[Brian Gaff (Sofa]

Yes and possibly aa high value resistor and a capacitor as well. Brian

[Dave Plowman (News]

Well remember the first SS relays I was aware of. Used to switch the 'on air' lights in a newly refurbished TV studio. Hardly a day existed where they all worked. ;-)

[Brian Gaff (Sofa]

The main issue with low current devices and relays is tarnishing I used to find. Some of the ones in cases were not really sealed very well and the contacts easily corroded due to the atmosphere. Speaker relays seem to suffer dirty contacts as well though and these switch at zero current but do need to take high currents in use which is when I find the channels go up and down till you clean their contacts. Sorry for the digression. I'm still not sure what the original query was attempting to do and why exactly. Brian

[Dave Liquorice]

Depnds on the application. I doubt a traditional relay would last long operating with a 7 Hz PWM signal to control the speed of a mains circulating pump. Might be a tad noisy (acoustically and electrically) as well. B-)

[The Natural Philosopher]

Pore Ole Dave. No, any person knowledgeable in electronics could tell you that .

Mechanical stuff wears out and arcs and sparks make it so., Solid state stuff suffers dopant migration, but under reasonable temperature conditions the lifetime will be 10-1000 times greater.

[Dave Liquorice]

No worries, I think everyone is confused about what you are trying to achieve.

What is the failure mode of these relay(s)? Omron are a reputable make with good specs and normally good life times.

OK follow that. Can't think what is going on to kill relays under the conditions expected by that description. The PIR isn't likely to switch often enough to wear out a relay in a time that makes replacing it an annoyance. Say 10 operations/night, 500,000 operations lifetime of relay = 137 years. Even at 100 ops/night that's 13 years! and the 500,000 could be on the low side, 5,000,000 might be nearer. B-)

Unfortunately that doesn't help a great deal. We need to know more about what the camera is expecting to find at its trigger input. What make/model camera are they?

[Martin Brown]

Lifetimes for relay based logic running a few daily tasks should be of the order of 30 or 40 years if properly designed. That is about how long my oil burner controller lasted before relays started to fail.

If they were not being driven exactly right then you were very lucky none of them caught fire. Though our loads were a lot more than a lamp. Modern ones are more tolerant but still not bomb proof.

[Adrian Caspersz]

Next time try giving a clue like the make or model number. It helps!

Googling "This is the cause or stimulus which defines when to trigger the Network Camera", this camera looks to be a vivotek_pz8121 or similar.

The camera input has a pull up to 12v, and shorting that to ground triggers it to record.

The 10A relay contacts are probably now some tarnished metal as the current is too low to keep them 'wet'[*], and probably a better choice will be one specified to switch signals rather than power.

Something like a reed relay with gold/silver plated contacts.

The coils of these tend to low voltage, so you will need a transformer like an an old mobile phone charger or maybe an LED driver transformer which usefully have terminals you can wire to, and maybe also use to switch on additional LED lighting?

You can use other components to get rid of the transformer, and use an optocoupler instead of the relay, but that might be taking a risk with electronics construction abilities and mains voltages.

• -

[Adrian Caspersz]

Ok, this is a cheap ready assembled pcb without a relay.

Should be OK to switch 12V, the EL817 chip used is rated to 35V at 50mA.

The UK eBay seller neatly covers himself with "Please be very careful with mains electricity - it can kill!"

[Paul]

Page 81 "This is the cause or stimulus which defines when to trigger the Network Camera"

11,570,114 bytes

Page 5 Shows "Digital Input" is 12V logic with a pullup inside the camera. Asserting a logic 0, requires pulling a very small (logic) current to ground

Page 81 Shows how the recording software can be programmed to use the Digital Input to collect a video stream.

You don't want a relay for this. That's just a logic input, needing hardly any drive capability at all.

You want an optoisolator capable of sinking that tiny logic current from a 12V level source. The optoisolator output transistor has some maximum voltage it can work with.

***********************************************************

An optoisolator looks like this. +----- resistor ------ +12V fed | inside camera to camera ----- resistor --+ +------------------------+ (design | | "Digital Input" <=== make this "active low" needed LED phototransistor Pin 3 here) | or photodarlington -----------------+ | Pin 4 Camera +------------------------+---------------------- ground ^^^^^^^^ double insulated barrier keeps mains off secondary side [some rated 5kV breakdown voltage]

You can use the ammeter on your multimeter, measure using the amps plug holes on the meter, and "amp" between Pin 3 and Pin 4, to see what sinking current is required of the optoisolator.

You can see in the discussion thread here, it's a target-rich environment, with all sorts of ideas for dangerous mains fun. They came up with more ideas than I did, so I leave it here.

Your PIR needs a resistive load, to make it work properly. There's likely a minimum light bulb it could run for example. You'd use a power resistor instead, which would last longer than a light bulb and not be a liability in the design (camera didn't record when you expected).

---- PIR ------+---- cct here ... --+ +------------------ Digital-In | | | power LED Photo resistor (opto transistor (keep PIR happy) isolator) | } | |

----------------+---------------------+ +------------------ Camera GND 6 pin or 8 pin opto Notice there is *no* connection between mains side and PhotoTransistor side!

Some people, when building circuit boards, leave optical hints on the board (white paint markings) to highlight where the lethal portion is and where the low voltage portion is. Allow sufficient spacing between components (creepage and clearance) to avoid any possibility of mains bridging to the secondary.

You want a Vero board with *no* copper on it for this. All connections via flying lead (insulated wire).

I would point you at a commercial module for this, if I could find one... No luck so far. Just don't know good keywords for it. Using obvious names isn't working.

In the Raspberrypi forum, they're not paying enough attention to mains transients, and some thought should be put into that aspect.

The most obvious circuits for this... draw too much power. That's why some trickery is required. I wanted to use an orange gumdrop (vitamine cap) for this, but it doesn't have enough microfarads. And I don't want to be pissing around with monster electrolytics that can puncture an eardrum if they arc over.

It would be nice if the PIR could drive a small mains transformer, as that would ease the LED driving process considerably. But transformers aren't cheap, and the last one I ordered took six weeks to show up. Back when they were manufactured locally, the process worked better.

I would not want the PIR to drive a wall wart SMPS, because of the inrush current those draw. I'm trying to craft something that the PIR "hardly notices".

LED efficiency has gone up considerably over the years, which is why some optos can get away with 1mA, whereas decades ago you needed 20mA to work them. The LED driving current has to be "smooth enough", to keep the phototransistor on the other side saturated. You don't want the Digital-In "humming at 50 cycles" :-) That would make the camera crazy. The circuit needs some smoothed-out DC for that LED.

Paul

[The Natural Philosopher]

The little opto isolator board I linked to should do exactly that

Just connect ground to ground and the output to the camera trigger. Leave the other output pin disconnected It should work!

Obviously a housing would be needed for outdoor work

[The Natural Philosopher]

Great minds...

Ah,. I hadnt got that far!

Of course